| Abstrakt: |
Background Recent efforts to prevent psychosis tried to identify at risk individuals in the community and in clinical settings. For this purpose, ultra-high risk for psychosis (UHR) and clinical high-risk state for psychosis (CHR-P) criteria were developed. The use of such criteria has proven not to be satisfactory, though, for conversion rates greatly vary across studies. The present study aims to assess whether cognitive patterns can be distinguished among a community sample of UHR subjects and controls, in order to increase specificity of UHR criteria. Methods Over 2,500 individuals aged 18–30 years were randomly interviewed in a household survey with the Prodromal Questionnaire (PQ). Those screened positive were recruited to the study's second phase, at the Institute of Psychiatry; 226 individuals agreed to participate, and their UHR status was assessed with the Structured Interview for Prodromal Syndromes (SIPS); 98 were regarded as UHR individuals, 124 as controls, and 4 as psychotic. Cognitive and childhood trauma data were available for 137 subjects (83 UHR and 54 controls), constituting our final sample. Cognition was assessed with the University of Pennsylvania Computerized Neuropsychological Testing, comprised of 11 tests that elicits 409 variables. Childhood adversity was evaluated with the Childhood Trauma Questionnaire, assessing 5 domains: physical abuse and neglect, emotional abuse and neglect, and sexual abuse. Latent profile analysis (LPA) was used to assess if there were discrete neuropsychological patterns across our sample–all 137 subjects were entered, regardless of UHR/control status. To account for local independency–a prerequisite of LPA model–out of the 409 neuropsychological variables, 24 main variables were selected. The following fit indices were used to assess the optimal number of latent classes: Akaike information criterion, sample-size adjusted Bayesian information criterion, bootstrapped likelihood ratio test, entropy, and the Lo-Mendell-Rubin's adjusted likelihood ratio test. LPA divided the sample into mutually exclusive neuropsychological patterns. Then, within each pattern, ANOVA with Bonferroni correction analyzed whether childhood trauma was associated with UHR status. SPSS and Mplus 8 for Mac were used for all analysis. Results LPA divided the sample into 6 cognitive patterns. Patterns 4 and 6 only had women, and pattern 2 had the lowest educational level. There were no significant differences in the distribution of UHRs and controls within each pattern. Pattern 1 and 2 subjects were characterized by lower visual attention, vigilance, and abstraction. Pattern 3 showed average results. Pattern 4 had lower word and facial memory, and lower visual object learning and memory. Pattern 5 had overall average results and better abstraction. Pattern 6 had lower emotion discrimination, facial memory, and higher perseverative errors. For pattern 3, UHRs compared to controls had significantly higher scores on physical abuse (1.75 vs 1.27, respectively, p=0.011) and emotional abuse (2.49 vs 1.70, respectively, p=0.000). For pattern 4, UHR compared to controls had significantly lower scores on physical neglect (1.17 versus 2.00, respectively, p=0.001). Discussion Our results showed that there is no specific cognitive profile related to UHR status in our populational sample. However, childhood trauma, a factor usually related to psychosis and risk for it, was only related to UHR in two specific cognitive profiles. Our study might shed light into the identification of different subgroups of at-risk individuals. Future studies should indicate whether these different cognitive profiles are related to distinct outcomes of the at-risk condition. [ABSTRACT FROM AUTHOR] |
